Ecological Indicators 69 (2016) 118125

Contents lists available at ScienceDirect

Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind

The renewable energy-growth nexus with carbon emissions and

technological innovation: Evidence from the Nordic countries
Manuchehr Irandoust
Halmstad University, School of Business, Engineering, and Science, PO Box 823, SE-30118 Halmstad, Sweden

a r t i c l e

i n f o

Article history:
Received 24 January 2016
Received in revised form 24 March 2016
Accepted 29 March 2016
Available online 22 April 2016
JEL classication:
Q2
Q3
Q4
Q5
C12
Keywords:
Renewable energy
CO2 emissions
Economic growth
Granger non-causality
Technological innovation

a b s t r a c t
Despite the fact that previous studies have extensively investigated the renewable energy-growth nexus,
those studies have not considered the role of technological innovation. This study examines the relationship between renewable energy consumption, technological innovation, economic growth, and CO2
emissions in the four Nordic countries by constructing a vector autoregression (VAR) model. On the basis
of a modied version of the Granger non-causality test, the results show a unidirectional causality running from renewable energy to CO2 emissions for Denmark and Finland and a bidirectional causality
between these variables for Sweden and Norway. The ndings also indicate a unidirectional causality
running from technological innovation to renewable energy and from growth to renewable energy for
the four Nordic countries. The results could not conrm any causality from renewable energy to growth.
Three policy implications are offered: (i) renewable energy improves environmental well-being, (ii) the
Nordic countries have very low energy intensities and high energy efciencies, and (iii) technological
innovation plays an effective role in the renewable energy-growth nexus.
2016 Elsevier Ltd. All rights reserved.

1. Introduction
As environmental issues, in particular climate change, become
increasingly important, energy consumption and production come
under intense scrutiny. The increasing magnitude of global energy
consumption and its rapid growth have severe environmental
implications. The global human population continues to grow, and
countries are continuing to develop, causing steep increases in the
demand for energy. At the current rate of increasing energy use,
energy demand is expected to increase 65% from 2004 levels by
2030 (Sims et al., 2007). However, the majority of energy currently
used globally is derived from non-renewable sources (Ristinen and
Krushaar, 2006; Sims et al., 2007).
Ever since the oil crisis of the early 1970s, the four Nordic
countries (Denmark, Finland, Norway, and Sweden) have invested
heavily in the search for alternative energy sources. Measures to
phase out the use of oil have proceeded smoothly.1 In 1970, oil
accounted for more than 70% of energy supplies in the Nordic

Tel.: +46 721631811.

E-mail addresses: manuchehr.eirandoust@hh.se, manuch.dost@gmail.com
1
All data in this section come from EuroStat and International Energy Agency
(IEA).
http://dx.doi.org/10.1016/j.ecolind.2016.03.051
1470-160X/ 2016 Elsevier Ltd. All rights reserved.

countries; by 2012, the gure was around 20%, mainly due to

the declining use of oil for residential heating. Thus, the Nordic
countries want to promote efcient and sustainable energy use and
a cost-effective energy supply that would facilitate the transition to
an ecologically sustainable society. These countries consume more
energy, yet their carbon emissions are low compared with those
of other countries. They release around 5 tonnes of carbon dioxide
per year into the atmosphere, compared with the EU average of
8 tonnes and the US average of 19 tonnes. The reason for this low
emissions rate is that most part of electricity in the Nordic countries
comes from nuclear power and hydroelectric power (neither of
which generates carbon emissions), cogeneration from combined
heat and power, and wind. In the transition to a sustainable society,
wave power may be an important technology in the future, but it
is still relatively undeveloped.
However, the growing concern about climate change has
emphasized the importance of renewable energy. It is widely
believed that renewable energy is virtually carbon free energy
sources that can serve as a potential solution to both energy security and climate change problems. The Nordic countries have made
a signicant investment in renewable energy sources and are now
regarded as the leaders in the area.
Renewable energy sources accounted for an 11.8% share of the
EU-28s gross inland energy consumption in 2013. The relative

M. Irandoust / Ecological Indicators 69 (2016) 118125

importance of renewable energy in gross inland consumption is relatively high in Denmark (24.2%), Finland (29.2%), Sweden (34.8%),
and Norway (37.4%). The Nordic electricity production is two thirds
renewable. This is due to the large amount of hydropower in these
countries (especially Norway and Sweden) but also to growing
sources of other renewable energies. Biomass is burned in combined heat and power plants across Finland and Sweden, while
Denmark has the highest share of wind power in the world. In addition to renewable energy, nuclear power in Sweden and Finland
means that the regions electricity is mostly CO2 -free.
Previous studies overall suggest that an increase in the use of
alternative energy is usually negatively related to the degree of
CO2 emissions for different economies, although the impact of
alternative energy on economic growth remains inconsistent (e.g.,
Varun and Bhat, 2009; Menyah and Wolde-Rufael, 2010; AlFarra
and Abu-Hijleh, 2012; Fadel et al., 2013; Lee, 2013; Sbia et al., 2014).
Recently, Lee (2013) has incorporated the international technological spillover effect in the nexus of clean energy, carbon emissions
and economic growth. He nds no relation between foreign direct
investment (FDI) and clean energy, and insignicant association
between FDI and carbon emissions for the group of G20. By the
same time, Tang and Tan (2013), and Fei et al. (2014) reveal a signicant relationship between electricity consumption, economic
growth, and technological innovation.2 Their results are in line with
the applicability of the endogenous growth theory on the energy
sector. Thus, one could argue that the endogenous growth theory
(the proxy of technological innovation) is more relevant than the
FDI spillover effect in examining the energy-growth nexus.
Thus, the aim of this study is to examine the causal linkage
between the renewable energy consumption, CO2 emissions, technological innovation in renewable energy sources, and economic
growth in order to nd out whether the same pattern as mentioned above exists for the Nordic countries. The study covers the
period 19752012 and by constructing a vector autoregression
(VAR) model developed by Toda and Yamamoto (1995) and Yamada
and Toda (1998), it is expected to improve the standard F-statistics
in the causality test procedure. To the best of our knowledge, this is
the rst study to investigate the causal linkage between renewable
energy and its related variables in the Nordic countries.
The reminder of the paper is organized as follows. Section 2 presents some stylized facts about energy sector in the
Nordic countries, Section 3 reviews theoretical considerations and
previous studies, Section 4 presents the data description and
methodology. Section 5 provides the empirical ndings. The last
section offers conclusions.

119

and Denmark. Consequently, oil today is used predominantly in the

transport sector.
In the past decades, it has been a steady growth in renewable
energy sources like biomass and wind. The amount of hydroelectric
power produced in the Nordic region has only experienced minor
growth and despite the increase in renewable and nuclear energy,
the absolute demand for fossil fuels has not changed dramatically
since 1971. This is due to an increase in the demand for energy
in the Nordic countries except for Denmark. The growth has been
especially strong in Norway and Finland. Population growth, a
higher share of single person households and economic growth
describe this development. The higher generation of low-carbon
energy has come in addition to, not instead of, fossil fuels. As a
forecasting for the year 2050, Nordic demand must drop in order
to meet the national climate targets, as outlined by the IEA and
Nordic Energy Research. Fossil fuels, which today make up 45% of
total primary energy supply, will have to be reduced to account for
only 16%, with a corresponding increase in the shares of nuclear
and renewable energy.
Energy intensity reduces as energy efciency in demand sectors like buildings and industry increases, but the overall structure
economy also affects the ratio. Countries characterized with signicant heavy industries will typically have a higher energy
intensity than countries earning GDP primarily from service sectors. Denmark has the lowest ratio which implies that there exists
energy efciency in production process and there is no energy
intensive industries. By comparison, Finland, Sweden, and Norway
have higher energy intensities due to manufacturing industries.
Countries dominated by primary and secondary sectors, such as
Denmark and Norway, have steadily decreased their energy intensities since 1990. Sweden and Finland have also followed this trend.
Even with respect to fossil-fuel intensity, the Nordic countries are
well ahead of the other countries in their fossil-fuel efciency
because of their high utilization of renewable energy. Fig. 1 shows
time plot of renewable energy consumption in terms of thousand
tonne in the Nordic countries from 1975 to 2012.
The total energy consumption in the Nordic countries is equal
to about 8% of energy consumption in the EU-27 in 2012. Industry
accounts for about a third of energy use in the Nordic countries on
average. Large hydroelectric potential and a rich endowment of raw
materials and natural resources such as wood and minerals have
played a signicant role in the development of energy-intensive
industry in the region. In Finland and Sweden, forest-based
industries like paper, pulp and timber are especially important.
Norway has signicant metal manufacturing based on historically
cheap and plentiful hydroelectricity. Norway has one of the worlds

2. Some stylized facts about energy sector in the Nordic

countries
Over a third of the Nordic regions energy supply originates
from renewable sources.3 The largest of these is biomass and
waste which are used to generate electricity, heat and transport
fuels. Renewable electricity in the region is also generated from
hydropower as well as growing share of wind power. With nuclear
power in Sweden and Finland, over half of the regions energy is
CO2 -free. Oil is still the largest single energy source, because of its
central role as a transport fuel. The 1970s oil crises led to a move
away from oil and toward alternative energy sources in power generation which is evident in an increase in the use of nuclear energy
in Sweden and Finland, as well as in a rise in the use of coal in Finland

2
They use aggregate data, i.e., number of patenting activities as a proxy for technological innovation.
3
The data in this section come from International Energy Agency (IEA), Nordic
Energy Research, and Nordic Center for Spatial Development.

Fig. 1. Renewable energy consumption in the Nordic countries (19752012).

120

M. Irandoust / Ecological Indicators 69 (2016) 118125

Fig. 2. CO2 emissions per capita in the Nordic countries.

Fig. 3. Real GDP per capita in the Nordic Countries (19752012).

highest electricity consumption per capita due in part to its aluminum industry. Denmark has not the same access to energy
resources as the other Nordic countries, reecting the small share of
industrial energy use and their low per-capita energy consumption.
The Nordic region has a relatively decarbonised electricity supply but greenhouse gas emissions are rather high due in part to the
cold climate and prevalence of energy-intensive industry. Danish
emissions are originated from its relatively carbon-intensive electricity production compared to other countries, but offset by its
lack of energy-intensive industry. Denmark has been among the
most successful countries at decoupling emissions and energy use
from economic growth. The Danish economy has grown by 78%
since 1980, while CO2 emissions have declined. This pattern indicates an increase in energy efciency, and increased use of wind
and biomass in electricity and heat production.
Finnish per capita emissions are among the highest in the region,
because of the use of fossil fuels in its electricity mix and its signicant industrial activity. Industry is the largest energy end-use sector
in Finland, with paper, pulp, metal and chemical being the main
industrial energy consumers. Norways relatively high per capita
energy consumption is offset by its renewable supply of electricity. Substantial petroleum extraction accounts for about 16% of the
countrys emissions. Swedens per capita greenhouse gas emissions
are the lowest in the Nordic region and they have been reduced substantially during the last decades, despite economic growth. This
trend has been outcome of two most important factors namely, the
introduction of nuclear power in the 1980s and a change-over from
oil to bio-fuels in the production of district heat. Fig. 2 shows time
plot of CO2 emissions per capita (in terms of tones) in the Nordic
countries between 1975 and 2012.
There exist signicant differences in energy use per capita in
the Nordic region. While all four countries share a relatively cold
climate, variation in industrial activity and energy resources lead
to remarkable differences in energy and electricity use per capita.
Norway has electricity-intensive industries, taking advantage of
abundant renewable sources of electricity, such as geothermal and
hydropower. Norway uses electricity in the heating of space and
water to a much larger extent than the other Nordic countries,
which have more developed district heating systems though underground water pipes, and is much more widespread in the Nordic
countries than other regions. Sweden and Finlands industrial sectors are made up by more heat intensive industries, such as the
paper pulp industry. Denmark meanwhile, has relatively little
energy-intensive industry. Although a total of 35% of the energy
supply and 65% of electricity production in the Nordic countries is
renewable.

This comes from the large amount of hydropower in Norway,

and Sweden, but also to growing sources of other renewable
sources. Biomass is burned in combined heat and power plants
across Finland and Sweden, while Denmark has the highest share
of wind power in the world. Beside renewable sources, nuclear
power in Sweden and Finland means that the regions electricity
is 85% CO2 -free. At almost two thirds, the Nordic region has one
of the highest shares of renewable electricity consumption in the
world. All four Nordic countries stand above the EU27 average, with
Norway having close to 100% renewable electricity consumption.
Generally speaking, the Nordic region has successfully decoupled GDP growth from greenhouse gas emissions in the last decade.
Decoupling economic growth from environmental damage has
been widely supported as a key step toward sustainability. This
can also be described as a reduction in the energy intensity of the
Nordic economies, a measure of the energy used to create one unit
of GDP. The region has also down a gradual decoupling of energy
use from greenhouse gas emissions, due to a decarbonisation of
energy production and efciency improvements.
The commitment of the Nordic countries for renewable energy
may be measured by the development of specic policy instruments to increase renewable energy production. Carbon taxes
(Norway), feed-in tariffs (Denmark), investment grants (Finland),
and electricity certicates (Sweden), serve as examples. Another
crucial measure is the support for research and development of
new energy technologies. All the Nordic countries except Norway
(spending 1.7% of GDP) spend more on R&D than EU average
of 2.0%. However, if the Norwegian R&D expenses were calculated per capita, spending would be higher due to Norways
extraordinarily high GDP. While R&D expenses in EU27 have constituted an unchanged proportion of GDP since the mid-1990s, the
R&D expenses of the Nordic countries have typically increased.
Figs. 3 and 4 indicate time plots of real GDP per capita and real
RD spending on energy sector in the Nordic countries from 1975 to
2012, respectively.
As a summary, the Nordic countries have achieved signicant
progress in the use of renewable sources of energy during the last
two decades. On average, the Nordic countries create electricity
from renewable sources at four times the level of the OECD
countries. However, there are considerable variations among
countries and regions mainly as a result of the availability of
natural resources. Norway almost produces their electricity from
renewable energy sources. While, in Norway hydropower stands
for almost 100% of all the electricity generation from renewable,
in Denmark produces approximately 30% of its electricity from
renewable sources of which approximately 64% is generated

M. Irandoust / Ecological Indicators 69 (2016) 118125

Fig. 4. RD spending on energy sector in the Nordic countries (1975201).

from wind power with the rest coming from solid biomass and
municipal waste. In Finland and Sweden biomass and hydropower
are the main sources of renewable energy, which combined stand
for approximately 35.5% in Finland and 55% in Sweden of the total
generation of electrical power. Hydropower in Sweden stands for
approximately 84.6% of the total electric power generated from
renewable sources while in Finland this gure is approximately
58%.
The Nordic countries have a strong position worldwide in energy
innovation due to strong national support for this sector. These
countries stand for more than 30% of the worlds market in the
production of wind energy technology. Innovation in bio-energy is
also strong with a share of almost 30% of all biomass-based generation of heat and power in the industrialized world and around 10%
of the total scientic knowledge production. Energy innovation is a
very important economic activity in the Nordic countries assuming
approximately 6% of total revenues and employment in the region
while exports of energy technology and equipment accounts for
approximately 59% of total industrial exports. Energy innovation
systems in the Nordic countries differ with respect to the energy
resources available, technology regimes, institutional structures
and the political commitments on energy and climate change.
3. Theoretical considerations and previous studies
Although empirical ndings regarding the relationship between
alternative energy, CO2 emissions and economic growth are generally mixed and contradictory, the relationship between the
variables can be classied into four hypotheses. First the growth
hypothesis which indicates that energy contributes to economic
growth both directly in the production process and/or indirectly as
a complement to labor and capital. In the Granger-causality sense,
the growth hypothesis is supported if the increase in energy consumption leads to an increase in output. The implication of such
a hypothesis is that, on the one hand, policies aimed at energy
conservation may potentially have a negative impact on economic
growth and, on the other hand, an increase in energy consumption might create a detrimental effect on economic growth due to
structural changes such as shifting from energy- intensive production toward less energy-intensive production. The negative impact
of energy consumption on growth may also stem from either
excessive energy consumption in unproductive industries, capacity
constraints or an inefcient energy supply.
Second the conservation hypothesis claims that energy conservation policies that are aimed at reducing CO2 emissions,

121

efciency improvement and waste management do not necessarily reduce growth and this kind of hypothesis is supported if an
increase in GDP Granger causes an increase in energy consumption.
It is also argued that an increase in GDP may also Granger cause a
decline in energy consumption when an economy is constrained
by political reasons, infrastructural hinders, and mismanagement
of resources which may create inefciencies and reduction in the
demand for goods and services, including energy consumption.
Third the feedback hypothesis asserts that there is an interdependent relationship between GDP and energy consumption
where each variable may behave as a complement to each other.
This implies that an increase (decrease) in energy consumption
leads to an increase (decrease) in GDP and the other way round
increase (decrease) in GDP may lead to increase (decrease) in
energy consumption. Therefore, feedback hypothesis is supported
by the evidence of bidirectional Granger causality between GDP and
energy consumption. Fourth, the neutrality hypothesis assumes
energy consumption to be relatively minor component of real GDP
and thus energy consumption should have no signicant impact on
economic growth. Similarly as in the case of the feedback hypothesis, energy conservation policies may have little or no affect on
GDP. The neutrality hypothesis is supported if there is no evidence
of Granger causality between energy consumption and real GDP.
There are another strand of the literature, i.e., studies examining the relationship between economic growth and environmental
pollution (e.g., Pao and Tsai, 2010). This kind of research tests
the validity of the so-called Environmental Kuznets Curve (EKC)
hypothesis. The EKC was rst discussed by Grossman and Krueger
(1995) who suggested that an inverted U shape relationship exists
between the level of pollution and economic growth. In other
words, the pollutant per capita and income per capita increase
together until a certain point of income is reached at which growth
of the pollutant attens and reverses. This implies that once a
certain level of income has been reached the further growth can
now be achieved without a proportional increase in emissions.
An important conclusion is that economic growth itself is the
solution to the environmental degradation. Although this hypothesis investigates the relationship between economic growth and
environmental pollution, it does not test the validity of the EKC
hypothesis and has a number of backwards such as: it considers renewable energy consumption instead of aggregate energy
consumption and thus it would make no sense to test the environmental degradation, and, furthermore, the methodology applied to
test the hypothesis is based on linear assumptions while testing the
EKC hypothesis requires modeling of non-linear relationship.
Another factor that previous studies (with a very few exception)
have neglected to incorporate into the energy-growth nexus is
technological innovation. The applicability of the endogenous
growth theory on the energy sector is important since technological innovation improves the efciency of energy generating
methods and the advancements in production techniques. Technological innovation is a better indicator than the technological
spillover effect via FDI since the latter cannot be realized without
local absorption capabilities. The endogenous growth theory is not
only applicable in explaining economic growth, but also a robust
framework in explaining the variation in renewable energy use.
Especially, technological innovation is a signicant independent
factor to reveal the variation in economic growth and renewable
energy use.
However, there are extensive studies that have focused on
the causal relationship between renewable energy consumption,
economic growth and CO2 emissions. For example, Soytas et al.
(2007) study the dynamic relationship between pollutant emissions, energy consumption and output in the US for the period
19602004. They nd that, in the long-run, economic growth
affects the use of energy and growth of pollution. Ang (2007),

122

M. Irandoust / Ecological Indicators 69 (2016) 118125

Soytas and Sari (2009), Hatzigeorgiou et al. (2011) examine the

relationship between economic growth, energy and environmental
pollution for different countries and periods. Generally speaking, studies report similar results with some narrow differences
observed in the results that might be due to the choice of different
countries, time period, and methodology.
Ang (2007) supports causality running from output growth to
CO2 emissions and energy consumption in France for the period
19602000. Hatzigeorgiou et al. (2011) report similar ndings
using a multivariate cointegration analysis to test the relationship
between CO2 emissions GDP and energy intensity in Greece during the period 19772007. They nd that there are unidirectional
causalities running from GDP to energy intensity and from GDP
to CO2 as well as bidirectional causality between CO2 emissions
and energy intensity. Soytas and Sari (2009) report a unidirectional
causality running from carbon emissions to energy consumption in
Turkey for the period between 1960 and 2000.
Sadorsky (2009) studies empirical model of renewable energy
consumption, CO2 emissions and oil prices for the G7 countries
from 1980 to 2005 using panel vector error correction model
(VECM). Panel cointegration techniques estimates show that in
long term, GDP per capita and emissions are two major factors
behind renewable energy growth per capita. Apergis et al. (2010)
examine the causal relationship between CO2 emissions, renewable energy, nuclear energy, and economic growth for a group of 19
developed and non-developed countries for the period 19842007.
They conclude that there is a long-run and positive relationship
between emissions and renewable energy consumption. Whereas,
the results from the panel Granger causality tests indicate that
renewable energy consumption does not reduce CO2 emissions in
the short-run. In the same way, Menyah and Wolde-Rufael (2010)
investigate the causal relationship between CO2 emissions, renewable and nuclear energy consumption and real GDP for the US for
the period 19602007. Their ndings show a unidirectional negative causality running from nuclear energy consumption to CO2
emissions and proves that nuclear energy consumption can help
to reduce CO2 emissions. In contrast, they nd no causality running from renewable energy consumption to CO2 emissions but
they nd a unidirectional causality running from CO2 emissions to
renewable energy consumption.
Vaona (2012) examines the relationship between energy consumption of (non) renewable energy sources and output using a
VAR model and a large data set in Italy over the period 18612000.
The results, in general, conrm bidirectional causality between
non-renewable energy consumption and output. This implies that
greater non-renewable energy consumption promotes economic
growth but an increase in level of output decreases the growth rate
of non-renewable energy consumption, which could possibly be
due to greater efciency in energy use. The study however could
not nd any evidence of causality between the renewable energy
consumption and output. Silva et al. (2011) examine the causal
relationship between GDP, CO2 , and renewable energy sources
(used for electricity generation only) in a sample of four countries
(Denmark, Portugal, Spain, US) during the period 19602004.

RECt

GDPt

CO
2t

Tang and Tan (2013) use patent data as a proxy for technical
innovation to study the causal relationship between CO2 emissions,
growth, and clean energy in Malaysia over the period 19702009.
The results based on ARDL approach shows that technological innovation is signicant in mitigating the usage of fossil fuels. Fei et al.
(2014) also use the number of patenting activities to explore the
causal relationship between technological innovation, CO2 emissions, growth, and clean energy in New Zealand and Norway over
the period 19712010. The results based on ARDL approach show
that technological innovation plays a signicant role in the clean
energy-growth nexus.
However, empirical studies seem to provide inconclusive and
mixed results. To the best of our knowledge, this is the rst study
to investigate causal linkage between renewable energy, economic
growth, technological innovation in renewable energy sources, and
CO2 emissions for the four Nordic countries (Denmark, Finland,
Norway, and Sweden) by using the Granger non-causality procedure developed by the Toda and Yamamoto (1995) and Yamada and
Toda (1998) which is expected to improve the standard F-statistics
in the causality test procedure.

4. Data and methodology

The variables used here are real GDP per capita, CO2 emissions
per capita (in terms of tones), real R&D spending on energy sector
(as a proxy for technological innovation), and renewable energy
consumption in terms of thousand toe. The data covers the period
19752012 and obtained from the World Bank database, the OECD
Statistics, and International Energy Agency database.4 The chosen period stems from the availability of data. All variables are
transformed in natural logarithms as it helps to minimize the uctuations in the data series.
Zapata and Rambaldi (1997) have shown that, in a regression
context, for determining whether some parameters of the model
are jointly zero, the traditional F-test is not valid when the variables
are integrated or cointegrated and the test statistics does not have
a standard distribution. This implies that the usual tests for exact
linear restrictions on the parameters (e.g. the Wald test) do not
have their usual asymptotic distributions if the data is integrated
or cointegrated.
To deal with this issue and to avoid the pre-testing distortions
associated with prior tests of non-stationarity and cointegration, the procedure proposed by Toda and Yamamoto (1995) and
Yamada and Toda (1998) is used here to ensure that the usual test
statistics for Granger causality have standard asymptotic distributions. They utilize a modied Wald test (MWald) for restrictions
on the parameters of a VAR (k), where k is the lag length in the
system. This test has an asymptotic chi-square distribution when
a VAR (k + dmax ) is estimated (where dmax is the maximal order of
integration suspected to occur in the system).
Following Toda and Yamamoto (1995) and Yamada and Toda
(1998), we can set up the following VAR(k + d(max)) model (see
also Appendix for more details):
RECti

k


GDPti
= B0 +
Bi

CO
2 ti

i=1

TINt
They conclude that for almost all countries except the US, the
increasing share of renewable energy sources have initially a negative impact on economic growth but positive impact on CO2
emissions reduction.

TINti

RECtj

dmax
 GDPtj
+
Bj

j=k+1 CO2 tj
TINtj

REC

GDP
+

CO2

(1)

TIN

4
Since data for R&D spending on energy sector is missing for Finland until 1989,
thus, from 1975 to 1989 total R&D spending was used as a proxy for technological
innovation. Data were obtained from Statistics Finland.

M. Irandoust / Ecological Indicators 69 (2016) 118125

123

Table 1
Test for Granger non-causality applying the Toda and Yamamoto modied Wald test.a
Country

Ho:

M-WALD

Denmark

REC does not Granger cause GDP

GDP does not Granger cause REC
REC does not Granger cause CO2
CO2 does not Granger cause REC
REC does not Granger cause TIN
TIN does not Granger cause REC

2.58
9.35*
14.73**
7.98
8.06
13.31*

Finland

REC does not Granger cause GDP

GDP does not Granger cause REC
REC does not Granger cause CO2
CO2 does not Granger cause REC
REC does not Granger cause TIN
TIN does not Granger cause REC

6.14
11.13*
15.79***
4.14
9.68
12.05**

Country

Ho:

M-WALD

Norway

REC does not Granger cause GDP

GDP does not Granger cause REC
REC does not Granger cause CO2
CO2 does not Granger cause REC
REC does not Granger cause TIN
TIN does not Granger cause REC

8.26
10.18***
17.40**
9.71*
6.55
14.35*

Sweden

REC does not Granger cause GDP

GDP does not Granger cause REC
REC does not Granger cause CO2
CO2 does not Granger cause REC
REC does not Granger cause TIN
TIN does not Granger cause REC

5.03
12.27**
21.24***
16.18*
7.87
19.49**

a
The order of k was chosen to be 2, 3, 2, and 2 for Sweden, Finland, Denmark, and Norway, respectively, based on the Schwarz criteria. The order of dmax was chosen to be
one for all countries, since it seems that each variable contains only one unit root at maximum.
*
Signicance level at 10%.
**
Signicance level at 5%.
***
Signicance level at 1%.

where REC is renewable energy consumption, GDP denotes per

capita growth, CO2 is emissions per capita, and TIN is technological
innovation or R&D spending on energy sector, B0 is a 4 1 intercept vector, B1 Bdmax are 4 4 matrices of coefcients and vec()
is white noise. Testing for Granger non-causality the general null
hypothesis is,
H0 : R = r,

(2)

where R is a (N (42 k + 4)) matrix of rank N and r is a (N 1) null

vector. N is the number of restrictions of the estimated coefcients
and = vec(B0 , . . ., Bk ). Testing the hypothesis of Granger noncausality from GDP growth to REC growth, (2) may be expressed
in terms of the coefcients as,
12
12
H0 : b12
1 = b2 = = bk = 0,

(3a)

where b12
are the coefcients of GDPt1 to GDPtk in the rst equai
tion of model (1). Evidence of causality from Growth to REC is
established by rejecting the null hypothesis. In a similar way, noncausality can be tested for the other direction. Evidence of causality
from REC growth to GDP growth is established by rejecting the
following null hypothesis, expressed in terms of the coefcients,
21
21
H0 : b21
1 = b2 = = bk = 0,

(3b)

are the coefcients of RECt1 to RECtk in the second

where b21
i
equation of model (1).
Rambaldi and Doran (1996) have shown that MWald methods for testing Granger non-causality can be computed by using a
seemingly unrelated regression (SUR). The main advantage of this
method is that it does not require information regarding the cointegration properties of the system, as argued by Zapata and Rambaldi
(1997). A SUR-type VAR model has a normal, standard limiting chisquare distribution and the usual lag selection procedure to the
system can be used even if there is no cointegration or if the stability and rank conditions are not fullled so long as the order of
integration of the process is not greater than true lag length of the
model (Yamada and Toda, 1998). Furthermore, VAR models can be
estimated using data in levels and testing for general restrictions
even if the process may present integration or cointegration of an
arbitrary order (Toda and Yamamoto, 1995).
To determine the maximum order of integration, it is necessary
to test for stationarity. The KPSS test is used to determine the integration order of each variable. Finally, a combination of multivariate
diagnostic tests is performed to check if the underlying statistical
assumptions are fullled.

5. Empirical results
The methodology discussed earlier to test for the Granger noncausality between renewable energy consumption, technological
innovation, economic growth, and CO2 emissions implies two steps.
First, the integration order of each variable needs to be determined.
The KPSS test shows that the variables are integrated of order one.
The results, not reported here, are available from the author on
request. The second step is to test for the Granger non-causality
using the Toda and Yamamoto procedure. The order of dmax is
chosen to be one for all countries in the sample since the variables contain only one unit root at maximum.5 The VAR(k + d(max))
model can be tested now for the Granger non-causality by using
MWald test statistic which has a standard asymptotic distribution.
The results of the Granger non-causality tests are shown in
Table 1. Although the degree of (statistically) signicance levels
differ across countries, they indicate that the null hypothesis of
Granger non-causality are rejected between renewable energy and
CO2 emissions, implying that the causal linkage is unidirectional
running from renewable energy to CO2 emissions in Denmark and
Finland and bidirectional between renewable energy and CO2 emissions in Sweden and Norway. The ndings also show that there is a
unidirectional causality running from technological innovation to
renewable energy and from growth to renewable energy in the four
Nordic countries. The results could not conrm any causality from
renewable energy to growth.
However, such evidence supports that the increasing use of
renewable energy sources helps to mitigate the carbon dioxide
emissions. In this case, policy makers should encourage a multilateral effort to promote renewable energy and to reduce CO2
emissions in the region. The Nordic countries possess the relatively
highest technological specialization index in hydro-related and
alternative energy. Drawing on the technological momentum of the
Nordic countries in developing and expanding the use of renewable
energy, the results show signicant unidirectional Granger noncausality between technological innovation and renewable energy
in these countries. Lack of causality from renewable energy to
growth implies that the renewable energy consumption is a relatively minor component of GDP and the countries have low energy
intensities and high energy efciencies.
Generally speaking, this study provided very strong evidence of
a causal relationship between renewable energy consumption and

5
Note that conventional theory applies to the asymptotic analysis of Walds test
even if variables are stationary.

124

M. Irandoust / Ecological Indicators 69 (2016) 118125

Table 2
DoornikHansen normality test of the residual.
Country

Ep-statistic

p-Value

Denmark
Finland
Norway
Sweden

8.29
7.07
14.10
11.71

0.44
0.36
0.11
0.23

well-being. Second, the Nordic countries have very low energy

intensities and high energy efciencies. Third, technological innovation or endogenous growth theory plays an effective role in the
energy-growth nexus.
Acknowledgements
The author would like to thank two anonymous referees for their
useful comments. Any remaining errors are my own.

Table 3
The LjungBox test for residual autocorrelation.
Country

L-B(12)

p-Value

Denmark
Finland
Norway
Sweden

33.13
26.48
31.36
24.19

0.10
0.49
0.21
0.57

CO2 emissions. All of the performed Granger non-causality tests

suggested the existence of bidirectional and unidirectional causality between the renewable energy consumption and CO2 emissions.
This nding is consistent with those of Silva et al. (2011) and implies
that renewable energy consumption helps to reduce CO2 emissions.
Furthermore, the tests indicated no evidence that would suggest
a causal relationship running from renewable energy use to economic growth, even though such a relationship was expected. This
can partly be explained due to the fact that the Nordic countries
have not only one of the lowest energy intensities in the world,
but also have a relatively low CO2 intensity, which allows these
countries to achieve high economic growth with very low energy
input and minimum CO2 emissions. Such results support the neutrality hypothesis which implies that energy is a relatively minor
component of real GDP and thus it should have no signicant
impact on economic growth.
It is well known that estimating a VAR with low degrees of
freedom could cause biases. Thus, the robustness of the result is
checked by considering the size of the VAR. For a xed order of
the VAR, for each country, growth variable is dropped so that the
VAR is left with remaining variables. The results, however, conrm
the earlier conclusion. Finally, a series of multivariate diagnostic
tests are performed to check that whether the underlying statistical assumptions of the VAR models are fullled, as shown by
Tables 2 and 3. All models are successful in dealing with the problem
of autocorrelation.
6. Conclusion
Although previous studies have extensively examined the causal
relationship between alternative energy, CO2 emissions and economic growth, those studies did not incorporate the effect of
technological innovation on the nexus of renewable energy-growth
nexus. This study includes the endogenous growth theory into the
renewable energy-growth nexus to provide a more robust framework regarding the causal relationship between the variables in the
four Nordic countries during the period 19752012 by constructing
a VAR model.
The results, based on the Granger non-causality procedure
developed by Toda and Yamamoto (1995) and Yamada and Toda
(1998), show that there are a unidirectional causality running
from renewable energy to CO2 emissions in Denmark and Finland
and a bidirectional causality between these variables in Sweden
and Norway. The ndings also show that there is a unidirectional
causality running from technological innovation to renewable
energy and from growth to renewable energy in the four Nordic
countries. The results could not conrm any causality from renewable energy to growth. There are three policy implications of
the ndings. First, renewable energy improves environmental

Appendix.
Here are the basic steps for the Toda and Yamamoto procedure:
1. Test each of the time-series to determine their order of integration. Ideally, this should involve using a test (such as the KPSS
test) for which the null is stationarity.
2. Let the maximum order of integration for the group of timeseries be m. So, if there are two time-series and one is found to
be I(1) and the other is I(2), then m = 2. If one is I(0) and the other
is I(1), then m = 1, etc.
3. Set up a VAR model in the levels of the data, regardless of the
orders of integration of the various time-series. Most importantly, you must not difference the data, no matter what you
found at Step 1.
4. Determine the appropriate maximum lag length for the variables
in the VAR, say p, using the usual methods. Specically, base the
choice of p on the usual information criteria, such as AIC, SIC.
Make sure that the VAR is well-specied. For example, ensure
that there is no serial correlation in the residuals. If necessary,
increase p until any autocorrelation issues are resolved.
5. Now take the preferred VAR model and add in m additional lags
of each of the variables into each of the equations.
6. Test for Granger non-causality as follows. For simplicity, suppose
that the VAR has two equations, one for X and one for Y. Test the
hypothesis that the coefcients of (only) the rst p lagged values
of X are zero in the Y equation, using a Wald test. Then do the
same thing for the coefcients of the lagged values of Y in the X
equation.
7. Its important that you dont include the coefcients for the
extra m lags when you perform the Wald tests. They are there
just to x up the asymptotics.
8. The Wald test statistics will be asymptotically chi-square distributed with p degrees of freedom, under the null.
9. Rejection of the null implies a rejection of Granger non-causality.
That is, a rejection supports the presence of Granger causality.
References
AlFarra, H.J., Abu-Hijleh, B., 2012. The potential role of nuclear energy in mitigating
CO2 emissions in the United Arab Emirates. Energy Policy 42, 272285.
Ang, J.B., 2007. CO2 emissions, energy consumption, and output in France. Energy
Policy 35 (10), 47724778.
Apergis, N., Payne, J.E., Menyah, K., Wolde-Rufael, Y., 2010. On the causal dynamics
between emissions, nuclear energy, renewable energy, and economic growth.
Ecol. Econ. 69 (11), 22552260.
Fadel, M.E., Rachid, G., El-Samra, R., Boutros, G.B., Hashisho, J., 2013. Emissions
reductions and economic implication of renewable energy market penetration
of power generation for residential consumption in the MENA region. Energy
Policy 52, 618627.
Fei, Q., Rasiah, R., Shen, L.J., 2014. The clean energy-growth nexus with CO2 emissions
and technological innovation in Norway and New Zealand. Energy Environ. 25
(8), 13231344.
Grossman, G.M., Krueger, A.B., 1995. Economic growth and the environment. Q. J.
Econ. 110, 353377.
Hatzigeorgiou, E., Polatidis, H., Haralambopoulos, D., 2011. CO2 emissions, GDP and
energy intensity: a multivariate cointegration and causality analysis for Greece,
19772007. Appl. Energy 88 (4), 13771385.
Lee, J.W., 2013. The contribution of foreign direct investment to clean energy use,
carbon emissions and economic growth. Energy Policy 55, 483489.

M. Irandoust / Ecological Indicators 69 (2016) 118125

Menyah, K., Wolde-Rufael, Y., 2010. CO2 emissions, nuclear energy, renewable
energy and economic growth in the US. Energy Policy 38 (6), 29112915.
Pao, H., Tsai, C., 2010. CO2 emissions, energy consumption and economic growth in
BRIC countries. Energy Policy 38 (12), 78507860.
Rambaldi, A.N., Doran, H.E., 1996. Testing for Granger non-causality in cointegrated
systems made easy, 88. Working Papers in Econometrics and Applied Statistics,
Department of Econometrics, University of New England, pp. 122.
Ristinen, R.A., Krushaar, J.J., 2006. Energy and the Environment. John Wiley and Sons,
New York.
Sadorsky, P., 2009. Renewable energy consumption, CO2 emissions and oil prices in
the G7 countries. Energy Econ. 31 (3), 456462.
Sbia, R., Shahbaz, M., Hamdi, H., 2014. A contribution of foreign direct investment,
clean energy, trade openness, carbon emissions and economic growth to energy
demand in UAE. Econ. Modell. 36, 191197.
Silva, S., Soares, I., Pinho, C., 2011. The Impact of Renewable Energy Sources on
Economic Growth and CO2 Emissions A SVAR Approach. Universidade do Porto,
Faculdade de Economia do Porto, 407 pp.
Soytas, U., Sari, R., 2009. Energy consumption, economic growth, and carbon
emissions: challenges faced by an EU candidate member. Ecol. Econ. 68 (6),
16671675.
Soytas, U., Sari, R., Ewing, B.T., 2007. Energy consumption, income, and carbon emissions in the United States. Ecol. Econ. 62 (34), 482489.

125

Sims, R.E.H., Schock, R.N., Adegbululgbe, A., Fenhann, J., Konstantinaviciute, I.,
Moomaw, W., Nimir, H.B., Schlamadinger, B., Torres-Martnez, J., Turner, C.,
Uchiyama, Y., Vuori, S.J.V., Wamukonya, N., Zhang, X., 2007. Energy Supply. In:
Metz, B., Davidson, O.R., Bosch, P.R., Dave, R., Meyer, L.A. (Eds.), Climate Change
2007: Mitigation of Climate Change. Contribution of Working Group III to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change.
Cambridge University Press, pp. 251322.
Tang, C.T., Tan, E.C., 2013. Exploring the nexus of electricity consumption, economic
growth, energy prices and technology innovation in Malaysia. Appl. Energy 104,
297305.
Toda, Y.H., Yamamoto, T., 1995. Statistical inference in vector autoregressions with
possibly integrated processes. J. Econ. 66, 225250.
Vaona, A., 2012. Granger non-causality tests between (non)renewable energy consumption and output in Italy since 1861: the (ir)relevance of structural breaks.
Energy Policy 45 (C), 226236.
Varun, P.R., Bhat, I.K., 2009. Energy, economics and environmental impacts of renewable energy systems. Renew. Sustain. Energy Rev. 13 (9), 27162721.
Yamada, H., Toda, Y.H., 1998. Inference in possibly integrated vector autoregressive
models: some nite evidence. J. Econ. 86, 5595.
Zapata, H.O., Rambaldi, A.N., 1997. Monte Carlo evidence on cointegration and causation. Oxf. Bull. Econ. Stat. 59 (2), 285298.